Public buy-in to Covid-19 policies like mask wearing and quarantining remains essential. More effective government communication would frame the potential outcomes from compliance or non-compliance – and target different parts of society with tailored messages.
Presenting scenarios is a key part of government communication strategies during the pandemic. Indeed, the chances that citizens will respond by adopting behaviours that prevent the spread of Covid-19 depend on such scenarios – which might be framed in terms of potential lives saved or potential deaths averted by protective measures. We know from research that the language used in framing a problem in the wider context of people’s lives, as well as their interpretation of the message, have significant effects on decision-making about health and risks.
There is now a vast body of evidence on public health campaigns showing that evoking fear or threats are not always effective in changing people’s behaviour with regard to their health habits. While fear appears to have some effects, especially if a significant and relevant threat is presented along with effective responses that seem easy to accomplish (White and Dillon, 2000), more often than not, fears and warnings can trigger people’s defensive mechanisms such as avoidance or denial (Ruiter et al, 2001).
This is because decisions are taken not just on the basis of the information provided but also involve complex social relationships in social contexts. For example, one study showed that HIV patients’ non-compliance with taking medicine had to do with the fact that many of the antiretroviral drugs had to be taken at meal times: since these were often social occasions with friends and family, and patients wanted to conceal the severity of their illness, they often delayed or skipped the dose (Jones, 2013).
Gain and loss framing in health information
In one of their seminal papers on the framing of decisions, Tversky and Kahneman (1981), the pioneers of ‘prospect theory’, introduced the ‘Asian disease problem’:
Imagine that the U.S. is preparing for the outbreak of an unusual Asian disease, which is expected to kill 600 people. Two alternative programs to combat the disease have been proposed. Assume that the exact scientific estimate of the consequences of the programs are as follows: If Program A is adopted, 200 people will be saved. If Program B is adopted, there is 1/3 probability that 600 people will be saved, and 2/3 probability that no people will be saved. [gain frame] If Program C is adopted 400 people will die. If Program D is adopted there is 1/3 probability that nobody will die, and 2/3 probability that 600 people will die. [loss frame]
A majority of respondents choose to be risk-averse when saving lives (‘gain framing’) but choose to take risks in avoiding deaths (‘loss framing’). One way to summarise the predictions of prospect theory is with the ‘fourfold pattern of risk preferences’: people are risk-averse over large expected gains and small expected losses; but risk-seeking over large expected losses and small expected gains (Markowitz, 1952; Scholten and Read, 2014).
Things get even more complicated in the domain of social decision-making. Most of the protective behaviours against coronavirus have ‘positive externalities’, protecting both the individual and others. For example, Batteux et al (2019) find that decisions made for others in the medical domain tend to be more risk-averse than for oneself, which in turn are known to be more risk-averse than one’s own financial decisions (Galizzi et al, 2016; Nebout et al, 2018).
Armstrong et al (2002) present subjects (aged 43 on average) with either death or survival rates for a disease and find that interest in preventative surgery is higher among those presented with survival rates. Nan (2012) studies gain versus loss framing on the decision to get the HPV vaccine among young adults: participants with a ‘present-focused mindset’ are more responsive to loss framing while those with a ‘future-oriented personality’ are not influenced by the framing of information.
In a meta-analysis of the effects of gain/loss framing on vaccination intentions, O’Keefe and Nan (2012) find no robust differences for people’s own behaviour but that parents are likelier to vaccinate their children when given loss-framed information. Pen?a and B?ban (2018) also find inconsistent effects, although Yu and Shen (2013) specifically examine the interaction of gain/loss framing with own/other consequences, and find that gains motivate individual behaviour whereas losses motivate behaviour on behalf of others.
A review by Rothman et al (2006) concludes that: ‘Gain-framed appeals are more effective when targeting behaviours that prevent the onset of disease, whereas loss-framed appeals are more effective when targeting behaviours that detect the presence of a disease’.
Covid-19 and framing
Coronavirus is likelier to affect older people. Therefore, the decisions taken by older people are likely to affect their own health to a greater extent than those of young people. Pornpattananangkul et al (2018) find that older people make less of a distinction between themselves and others when taking financial risks – that is, the consequences of risk happening to other people is more likely to change the behaviour of the young.
Initial evidence from the Covid-19 pandemic and the public health responses has yielded a number of reasons to believe that gain/loss framing affects the prophylactic measures that citizens take to protect themselves and others from the infection. Akesson et al (2020) document lower willingness to comply with protective measures among survey respondents presented with higher death estimates – an outcome that they call the ‘fatalism effect’.
Citizens’ willingness to comply with mandates to engage in protective behaviours may also depend on their confidence in the government and fellow citizens (Fetzer et al, 2020). The UK, the United States and Italy feature widespread perceptions that their governments are not doing enough to prevent the spread of the virus.
Framing may affect expectations about the number of other people complying, which could thereby influence behaviour where people are reciprocal (Jehiel, 2005; Dufwenberg et al, 2011). Müller and Rau (2020) find compliance with Covid-19 distancing policies to be correlated with social, risk and time preferences at the individual level, while Alfaro et al (2020) show that regional social preference measures fit infection rates.
Communication of Covid-19 social distancing measures may also affect compliance by conveying information about the likely costs to individuals of social distancing. Briscese et al (2020) find that compliance in a representative sample of Italian respondents is lower among those who expected the measures to last for a shorter time. It must be emphasised that people’s ability to engage in protective behaviours depends on many other factors, such as their ability to work from home (Chiou and Tucker, 2020).
Controlling for the latter, Biroli et al (2020) investigate the possibility of using gain/loss framed messages to encourage differential protective behaviours during Covid-19. They use a survey that presents respondents with projections of hypothetical outcomes where consequences of mass adherence to social distancing measures are randomly presented in terms of either lives lost or deaths averted.
They find that older respondents are more likely to report intentions to adhere to safety measures when the projection associated with adherence is presented in terms of lives saved; this effect does not hold for younger respondents. They also find differential adherence patterns to protective measures by how ‘socially oriented’ respondents are (that is, whether they are willing to cooperate for others’ benefit); and a generally higher number of protective behaviours reported by women than men.
The lessons are many: information framing is very important and differential groups need to be targeted differently in order to elicit the highest possible compliance. Although they might appeal to governments for reputational reasons, blanket campaigns are unlikely to be the most effective tool for combating Covid-19.
As the crisis moves into its long tail, the importance of measures like mask wearing and requests to isolate will continue to be important in preventing new flare-ups. Public buy-in to these policies is therefore essential. Messaging may improve public compliance with these measures by framing the potential outcomes. Targeting different populations with tailored messages can increase effective communication.
Where can I find out more?
- Framing the predicted impacts of COVID-19 prophylactic measures in terms of lives saved rather than deaths is more effective for older people: Study of compliance with public health messages about Covid-19 prevention in Italy, the UK and the United States by Pietro Biroli, Steven Bosworth, Marina Della Giusta, Amalia Di Girolamo, Sylvia Jaworska and Jeremy Vollen.
- Compliance with social distancing during the Covid-19 crisis: Paola Giuliano and Imran Rasul draw on real-time data collected across many different countries to document important drivers of compliance with social distancing.
- Civic capital and social distancing: evidence from Italians’ response to Covid-19: Ruben Durante and colleagues study the relationship between civic capital and social distancing across Italian provinces.